Merge branch 'main' into dev

colab/GPU.ipynb

@@ -98,7 +98,7 @@
         "!git clone https://github.com/Cohee1207/tts_samples\n",
         "!npm install -g localtunnel\n",
         "!pip install -r requirements-complete.txt\n",
-        "!pip install tensorflow==2.11\n",
+        "!pip install tensorflow==2.12\n",
         "\n",
         "\n",
         "cmd = f\"python server.py {' '.join(params)}\"\n",

docker/docker-compose.yml

@@ -1,12 +1,13 @@
 version: "3"
 services:
-  tavernai:
+  sillytavern:
     build: ..
-    container_name: tavernai
-    hostname: tavernai
-    image: tavernai/tavernai:latest
+    container_name: sillytavern
+    hostname: sillytavern
+    image: cohee1207/sillytavern:latest
     ports:
       - "8000:8000"
     volumes:
       - "./config:/home/node/app/config"
-    restart: unless-stopped
+      - "./config.conf:/home/node/app/config.conf"
+    restart: unless-stopped

public/scripts/openai.js

@@ -562,13 +562,19 @@ async function sendOpenAIRequest(type, openai_msgs_tosend, signal) {
             const decoder = new TextDecoder();
             const reader = response.body.getReader();
             let getMessage = "";
+            let messageBuffer = "";
             while (true) {
                 const { done, value } = await reader.read();
                 let response = decoder.decode(value);
 
                 tryParseStreamingError(response);
-
-                let eventList = response.split("\n");
+                
+                // ReadableStream's buffer is not guaranteed to contain full SSE messages as they arrive in chunks
+                // We need to buffer chunks until we have one or more full messages (separated by double newlines)
+                messageBuffer += response;
+                let eventList = messageBuffer.split("\n\n");
+                // Last element will be an empty string or a leftover partial message
+                messageBuffer = eventList.pop();
 
                 for (let event of eventList) {
                     if (!event.startsWith("data"))

readme.md

@@ -65,6 +65,8 @@ Get in touch with the developers directly:
   * Character emotional expressions
   * Auto-Summary of the chat history
   * Sending images to chat, and the AI interpreting the content.
+  * Stable Diffusion image generation (5 chat-related presets plus 'free mode')
+  * Text-to-speech for AI response messages (via ElevenLabs, Silero, or the OS's System TTS)
 
 ## UI Extensions 🚀
 
@@ -76,6 +78,8 @@ Get in touch with the developers directly:
 | D&D Dice | A set of 7 classic D&D dice for all your dice rolling needs.<br><br>*I used to roll the dice.<br>Feel the fear in my enemies' eyes* | None | <img style="max-width:200px" alt="image" src="https://user-images.githubusercontent.com/18619528/226199925-a066c6fc-745e-4a2b-9203-1cbffa481b14.png"> |
 | Author's Note | Built-in extension that allows you to append notes that will be added to the context and steer the story and character in a specific direction. Because it's sent after the character description, it has a lot of weight. Thanks Ali឵#2222 for pitching the idea! | None | ![image](https://user-images.githubusercontent.com/128647114/230311637-d809cd9b-af66-4dd1-a310-7a27e847c011.png) |
 | Character Backgrounds | Built-in extension to assign unique backgrounds to specific chats or groups. | None | <img style="max-width:200px" alt="image" src="https://user-images.githubusercontent.com/18619528/233494454-bfa7c9c7-4faa-4d97-9c69-628fd96edd92.png"> |
+| Stable Diffusion | Use local of cloud-based Stable Diffusion webUI API to generate images. 5 presets included ('you', 'your face', 'me', 'the story', and 'the last message'. Free mode also supported via `/sd (anything_here_)` command in the chat input bar. Most common StableDiffusion generation settings are customizable within the SillyTavern UI. | None | <img style="max-width:200px" alt="image" src="https://files.catbox.moe/ppata8.png"> |
+| Text-to-Speech | AI-generated voice will read back character messages on demand, or automatically read new messages they arrive. Supports ElevenLabs, Silero, and your device's TTS service. | None | <img style="max-width:200px" alt="image" src="https://files.catbox.moe/o3wxkk.png"> |
 
 ## UI/CSS/Quality of Life tweaks by RossAscends